Method for removing ring artifacts from tomograms produced with the aid of a computed tomography unit

A method and a computed tomography unit are disclosed, that render it possible in a simple way to remove ring artifacts from tomograms Ik, particularly in the case of a fast feed of the recording region per revolution of the recording system of the computed tomography unit by calculating a ring artifact image Ik for each tomogram Ik. In this procedure, temporary ring artifact images Rtk are firstly calculated for each tomogram Ik, and subsequently the final ring artifact image Rk is formed for the purpose of correcting the respective tomogram Ik by averaging over at least a portion Tsub of the temporary ring artifact images Rtk, the temporary ring artifact images Rtk being rotated before averaging in such a way that the ring artifacts present in the temporary artifact images Rtk substantially coincide and their position substantially corresponds to the ring artifacts present in the tomogram Ik.

Skip to: Description  ·  Claims  ·  References Cited  · Patent History  ·  Patent History
Description

The present application hereby claims priority under 35 U.S.C. §119 on German patent application number DE 10 2004 044 698.9 filed Sep. 15, 2004, the entire contents of which is hereby incorporated herein by reference.

FIELD

The invention generally relates to a method for removing ring artifacts from tomograms produced with the aid of a computed tomography unit, and/or to a computed tomography unit, such as one operating according to such a method for example.

BACKGROUND

The recording system of a computed tomography unit which is known, for example, from EP 0 819 406 A1 includes an X-ray machine and a multirow detector. An X-ray machine and detector are arranged lying opposite one another on a rotary frame such that an object volume in the form of a spiral scan can be scanned during rotation of the rotary frame about a rotation axis and during a continuous feed of the object, which is supported on a bearing frame, in the direction of the rotation axis. A tomogram or volumetric image is reconstructed on the basis of the X-ray images acquired by the detector during the spiral scan from different projection directions.

Ring artifacts that are centered about a center of rotation, identifiable in the tomogram, of the recording system of the computed tomography unit cause signal errors of individual detector elements of the multirow detector. DE 198 35 451 A1 describes a method for removing ring artifacts in tomograms produced with the aid of a computed tomography unit. The known method includes method steps for calculating a correction image that substantially includes only the ring artifacts present in the tomogram.

The tomogram is subjected to at least one instance of median filtering for this purpose. An instance of low-pass filtering following thereupon serves the purpose of suppressing interfering noise components and noise structures. The low-pass filtering is performed along circular arcs about the center of rotation in the image. In the case of the known method, a result image in which the ring artifacts are suppressed is produced by a subsequent subtraction of the low-pass-filtered correction image from the input image, it being necessary, however, for the ring artifacts to have a certain minimum length for a successful suppression.

The length of a ring artifact or of a ring segment representing a ring artifact is determined by the speed with which that recording region of an object which is to be scanned is displaced in the direction of the rotation axis during each revolution of the recording system. The ring segment becomes shorter in this case the quicker the feed of the recording region to be scanned. However, ring segments of a short length such as are produced during rapid scanning of the recording region are not acquired or suppressed at all in tomograms in the case of the known method, or only unsatisfactorily so.

SUMMARY

An object of at least one embodiment of the present invention is to specify a method with the aid of which an improved removal of ring artifacts in tomograms is possible in a simple way.

An object may be achieved by a method for removing ring artifacts from tomograms, and/or by way of a computed tomography unit, such as one operating according to such a method for example.

According to at least one embodiment of the invention, a method for removing ring artifacts includes reconstructing tomograms Ik at recording positions (k=1, . . . , N) that are substantially equidistant in the direction of a rotation axis of the computed tomography unit, and calculating a temporary ring artifact image Rtk for each tomogram Ik which has at least one ring artifact present in the respective tomogram Ik. The method further includes method steps for calculating a ring artifact image Rk for each tomogram, the respective ring artifact image Rk being formed by averaging over at least a portion of the temporary ring artifact images Rtk, and the temporary ring artifact images Rtk being rotated before averaging in such a way that the ring artifacts present in the temporary artifact images Rtk substantially coincide and their position substantially corresponds to the ring artifacts present in the tomogram Ik. In addition, a method step is included for subtracting the respective ring artifact image Rk from the respective tomogram Ik, at least one result image Ek being produced in which the ring artifact is removed.

At least one embodiment of the invention proceeds from the fundamental finding that ring artifacts in consecutive tomograms that are acquired substantially at equidistant recording positions and under the same scanning conditions are of the same shape but are rotated by an angle to one another about a respective center of rotation of the tomogram. In addition to the ring artifacts present in a respective tomogram, the temporary ring artifact images that can be calculated for each tomogram can also have additional image perturbations such as pixel noise, for example. The correction of the ring artifacts in a tomogram is not performed, as previously, on the basis of an individual image, but on the basis of an averaging over a portion of the temporary ring artifact images obtained at various recording positions, such that the image perturbations present in the respective temporary ring artifact image are suppressed by a low-pass effect. In order for the ring artifacts present in the temporary ring artifact images to coincide, the temporary ring artifact images are rotated with reference to the tomogram to be corrected.

By contrast with the known method, at least one embodiment of the invention therefore includes low-pass filtering in the time domain on the basis of images that have been acquired at different recording positions and/or at different times. In addition, interfering noise structures and interfering pixel noise are removed independently of the length of a ring artifact that is present.

According to an advantageous refinement of at least one embodiment of the invention, calculation of the temporary ring artifact image may be performed using the following method steps:

  • b1) masking bone and air fractions in the respective tomogram Ik such that a masked image Nk is respectively produced for each tomogram Ik,
  • b2) subjecting the respective masked image Nk to high-pass filtering in the radial direction relative to the center of rotation in the masked image Nk such that a high-pass-filtered image Hk is produced for each masked image Nk, and
  • b3) artifact threshold value formation in the respective high-pass-filtered image Hk with a negative artifact threshold and a positive artifact threshold, so as to produce for each high-pass-filtered image Hk the temporary ring artifact image Rtk that has the ring artifacts present in the respective tomogram Ik.

Masking of bone and air fractions may be performed advantageously by way of the following method steps:

  • b11) all the values greater than an upper threshold SWO are set equal to SWO, and
  • b2) all the image values smaller than a lower threshold SWU are set equal to SWU such that a masked image Nk is produced.

The high-pass filtering of the respective masked image may further include, for example, the following method steps:

  • b21) carrying out median filtering in radial directions, running through the center of rotation (Dz), in the masked image Nk such that a median-filtered image Mk is produced, and
  • b22) subtracting the median-filtered image Mk from the tomogram Ik such that a high-pass-filtered image Hk is produced.

The temporary ring artifact images Rtk contain not only ring artifacts, but also noise structures of object parts that have been falsely recognized as artifacts. These noise structures can preferably be eliminated by means of low-pass filtering in the respective temporary ring artifact image Rtk in the azimuth direction along at least one circular segment. In this case, the circular segments correspond to a part of a circle seated at the center of rotation, the circular segments covering, for example, an angular range of 10 degrees of a complete circle and being matched to the respective length of the ring artifact.

On the basis of the temporary ring artifact images Rtk thus determined, the ring artifact image Rk used for correcting the ring artifacts is preferably calculated for the respective input image using the following rule:

R k = l = - T + T w l · Rot l · δ ( Rt l + k ) ,
where l is an index traversing the values −T and +T, w1 is a weighting factor dependent on the index l, δ is the angular spacing of ring artifacts between two neighboring tomograms, and Rot is a rotation operator that rotates the temporary ring artifact image Rt1+k by the angle 1*δ about the center of rotation, the following relationships holding:

T sub = 2 * T + 1 , l = - T T w l = 1 and δ = 2 * π * d / V ,
and d corresponding to a difference between the neighboring recording positions and V corresponding to feed of the recording system in the direction of the system axis.

The weighting factors w1, which are each multiplied in accordance with the abovenamed calculation rule by the respective rotated temporary ring artifact image Rtk, take the following form:
w1=1/Tsub.

However, it is also possible to use any other desired weighting factors such that low-pass filtering can be performed with the aid of a characteristic matched to the situation. All that is essential is that the sum of all the coefficients yields 1.

Instead of low-pass filtering in the respective temporary ring artifact image Rtk, it is also possible to carry out low-pass filtering of the calculated final ring artifact image Rk. The low-pass filtering in the ring artifact image Rk in the azimuth direction is performed, again preferably, along at least one circular segment that corresponds to a part of a circle seated at the center of rotation and which has a circular segment length matched to the ring artifact. The circular segment length can, for example, cover an angular range of 10 degrees of a complete circle.

In an advantageous refinement of at least one embodiment of the invention, at least one method step is carried out in polar coordinates with an origin of coordinates seated at the center of rotation of the image.

According to at least one embodiment of the invention, a computed tomography unit is configured in such a way that it is possible to execute the method according to at least one embodiment of the invention for removing ring artifacts.

BRIEF DESCRIPTION OF THE DRAWINGS

Example embodiments and further advantageous refinements of the invention are illustrated in the following schematic drawings, in which:

FIG. 1 shows the essential parts of a computed tomography unit in an illustration that is partially like a block diagram and partially in perspective,

FIG. 2 shows, in outline form, a sequence of a method according to the invention of removing ring artifacts from tomograms with the aid of a computed tomography unit in accordance with FIG. 1, and

FIG. 3 shows a calculation of a ring artifact image on the basis of 3 temporary ring artifact images.

DETAILED DESCRIPTION OF THE EXAMPLE EMBODIMENTS

A computed tomography unit is shown in FIG. 1 in an illustration that is partially perspective and partially in the form of a block diagram. The computed tomography unit essentially includes a recording system 1, 2 having an X-ray machine 1 and a detector 2 that has detector elements 3 arranged in a detector array to form columns and rows, a computing device 5 for carrying out the method according to at least one embodiment of the invention for the purpose of removing ring artifacts, and a display unit 6 for displaying the result image Ek free from ring artifacts.

The X-ray machine 1 and the detector 2 are fitted opposite one another on a rotary frame (not illustrated) in such a way that during operation of the computed tomography unit an X-ray beam emanating from a focus F of the X-ray machine 1 and bounded by marginal rays 7 impinges on the detector 2.

The rotary frame can be set rotating about a rotation axis Da by way of a drive device (not illustrated). In this case, the drive axis Da is parallel to the z-axis of a rectangular three-dimensional coordinate system illustrated in FIG. 1. X-ray pictures can be prepared in this way from different projection directions or positions of angle of rotation for a recording region of, for example, a patient (not illustrated) supported on a patient table 4, for the purpose of reconstructing a number of tomograms Ik. The recording region is scanned in this procedure by way of a continuous feed of the patient couch 4 in the direction of the z-axis and in the form of a spiral scan 8 in the case of rotation of the recording system 1, 2 about the recording region to be examined.

Because of the spiral type of scanning, defectively operating or defective detector elements 3 of the detector 2 cause ring artifacts Ik in the respective tomogram Ik, which are visible as ring segments and are centered in the image with reference to the center of rotation. The length of the ring segments is a function of the speed with which the recording region is displaced in the direction of the z-axis during scanning per revolution of the recording system 1, 2. In principle, the ring segments are shorter the more rapid the feed of the recording region. For example, it is possible to conceive lengths of ring segments that only have an angular coverage of 10 degrees of a complete circle in the case of rapid scanning.

In conjunction with FIG. 3, FIG. 2 shows the sequence of the method according to at least one embodiment of the invention for the purpose of removing ring artifacts from tomograms, in the form of a flowchart. To simplify the illustration, the method is explained by way of example on the basis of 3 tomograms I0, I1, I2 at the recording position k=1.

The method of at least one example embodiment include the following steps, which are explained in detail below:

First Method Step A:

Reconstructing N, in this example 3 tomograms Ik (k=0, . . . 2) at recording positions 0, 1, 2 that are substantially equidistant in the direction of the rotation axis Da.

Second Method Step B:

Calculating a temporary ring artifact image Rt0 or Rt1 or Rt2 for each tomogram I0 or I1 or I2 that has at least one ring artifact S0 or S1 or S2 present in the respective tomogram I0 or I1 or I2.

Third Method Step C:

Calculating a ring artifact image R1 for the tomogram I1, the ring artifact image R1 being formed by averaging over the three temporary ring artifact images Rt0, Rt1, Rt2, and the temporary ring artifact images Rt0, Rt1, Rt2 being rotated before averaging in such a way that the ring artifacts S0 or S1 or S2 present in the temporary ring artifact images Rt0, Rt1, Rt2 substantially coincide and their position substantially corresponds to the ring artifact S1 present in the tomogram I1.

Fourth Method Step D:

Subtracting the ring artifact image R1 from the tomogram I1, in each case one result image E1 being produced in which the ring artifact S1 is removed.

Scanning a recording region, for example by way of the computed tomography unit shown in FIG. 1, is performed in such a way that tomograms I0 or I1 or I2 are reconstructed in the first method step A at substantially equidistant recording positions on the basis of the raw data 3 produced during scanning. Each tomogram I0 or I1 or I2 incldues M*M pixels arranged, for example, in a pixel matrix. Each pixel has a pixel value proportional to an x-radiation acquired by a detector element 3. Each pixel value can be represented in this case with the aid of a bit depth established by the detector 2. Given a bit depth of 12 bits, for example, the pixel values are in a range of between 0 and 4095.

In the present example, the tomograms I0, I1, I2 are to have at least in each case one ring artifact S0 or S1 or S2 that is caused by a defective detector element 3 of the detector 2 during spiral scanning of a recording region. In the present example, it is assumed, furthermore, for the purpose of simplified explanation of the method according to at least one embodiment of the invention that the feed of the recording region in the direction of the rotation axis Da is constant for each revolution of the recording system 1, 2. The ring artifacts S0, S1, S2 in the tomograms I0, I1, I2 reconstructed substantially at equidistant recording positions 0, 1, 2 are then visible as a ring segment that is rotated by the same angle δ in each case between neighboring tomograms I0, I1 or I1, I2.

After the N tomograms I0, I1, I2 have been reconstructed, a temporary ring artifact image Rt0 or Rt1 or Rt2 that substantially includes only the ring artifact S0 or S1 or S2 is calculated in the second method step B for each tomogram I0 or I1 or I2.

During masking B1 of the tomogram I0 or I1 or I2, the pixel values relating thereto are firstly subjected to threshold value formation SW with an upper threshold SWO and a lower threshold SWU such that the pixel values of image areas of bone and air fractions are limited, and are not falsely detected as ring artifacts during subsequent calculation. The threshold value formation SW is executed in such a way that pixel values that are greater than the upper threshold SWO and could originate from bone fractions are set equal to SWO, while pixel values that are smaller than the lower threshold SWU and could originate from air or air inclusions are set equal to SWU. The upper threshold SWO and the lower threshold SWU depend essentially on a set x-radiation, the dynamics of the detector 2 used and on the maximum attenuation value that can be observed, and can be determined empirically. The masked image N0 or N1 or N2 obtained in each case from this method step has M*M pixels in turn.

Subsequently, high-pass filtering B2 is carried out in the masked image N0 or N1 or N2 in the direction of a center of rotation Dz, imaged in the masked image N0 or N1 or N2, of the recording system 1, 2. The high-pass filtering B2 is preferably undertaken in this case by means of median filtering B21 along a multiplicity of straight lines running through the center of rotation Dz. The straight lines cover the pixel matrix in such a way that each pixel lies on such a straight line.

A median filter used for median filtering B21 can have, for example, 2*A1+1 (for example A1=3) interpolation points that are respectively arranged symmetrically relative to the pixel respectively to be processed. A spacing present between the interpolation points is selected such that approximately half a line width of the ring artifact S0 or S1 or S2 is covered by the median filter. Such dimensioning of the median filter ensures the detection of the ring artifact S0 or S1 or S2 while other image structures that are larger are advantageously suppressed.

By subtracting B22 a median-filtered image M0 or M1 or M2 obtained in this way from the associated tomogram I0 or I1 or I2, a high-pass-filtered image H0 or H1 or H2 is thus produced that is subjected to an artifact threshold value formation B3 with a lower artifact threshold and an upper artifact threshold. The artifact threshold formation B3 serves the purpose of suppressing image perturbations still possibly present in the image and which can occur in the region of bone edges, for example. Pixels in the high-pass-filtered image H0 or H1 or H2 that have a pixel value greater than the upper artifact threshold or smaller than the lower artifact threshold are identified as perturbations that have falsely been detected as ring artifacts. Such pixels identified as perturbations are limited with reference to their pixel value to the lower or upper artifact threshold such that a temporary ring artifact image Rt0 or Rt1 or Rt2 in which the ring artifact S0 or S1 or S2 is present is produced as a result.

Subsequent to the calculation of the temporary ring artifact image Rt0 or Rt1 or Rt2 relative to the tomogram I0 or I1 or I2, low-pass filtering T1 is undertaken in the image in order to suppress perturbing noise components. The low-pass filtering T1 is performed for each pixel in the temporary ring artifact image Rt0 or Rt1 or Rt2 in the azimuth direction along a multiplicity of circular segments that correspond in each case to a part of a circle seated at the center of rotation of the image, and which have a circular segment length matched to the ring artifact S0 or S1 or S2.

The circular segments cover the pixel matrix in such a way that each pixel lies on such a circular segment. For the sake of clarity, only one circular segment K0 or K1 or K2 is depicted in each case in the temporary ring artifact images in FIG. 3. The low-pass filter T1 can have, for example, 2*A2+1 (for example A2=3) interpolation points that are respectively arranged symmetrically relative to the pixels respectively to be processed along the circular segment K0 or K1 or K2.

A spacing present between the interpolation points is selected such that the low-pass filter T1 covers approximately the length of the ring artifact S0 or S1 or S2. Such a dimensioning of the low-pass filter T1 ensures an adequately effective suppression of high-frequency noise components in the image, while the structures of the ring artifact S0 or S1 or S2 are retained in the image. The low-pass filtering T1 can be implemented, for example, by simple averaging of the pixel values present at the interpolation points. However, it is also possible to conceive other types of low-pass filtering with a transmission characteristic or with another weighting of the interpolation points.

The second method step B is carried out for each tomogram I0 or I1 or I2 such that a temporary ring artifact image Rt0 or Rt1 or Rt2 is produced for each tomogram.

In the example shown in FIG. 2 and FIG. 3, the calculation of a ring artifact image R1 for the tomogram I1 is performed in a third method step C, the ring artifact image R1 being formed by averaging over the previously determined temporary ring artifact images Rt0, Rt1, Rt2, and the temporary ring artifact images Rt0, Rt1, Rt2 being rotated before the averaging in such a way that the ring artifacts S0 or S1 or S2 present in the temporary ring artifact images Rt0, Rt1, Rt2 substantially coincide and their position substantially corresponds to the ring artifact S1 present in the tomogram I1.

A calculation of the ring artifact image R1 at the recording position 1 is shown by way of example in FIG. 3 on the basis of the 3 temporary ring artifact images Rt0, Rt1, Rt2. The temporary ring artifact images Rt0, Rt1, Rt2 respectively contain the ring artifact S0 or S1 or S2, the ring artifacts S0, S1, S2 having the same form. Because of the spiral scanning, the ring artifacts of neighboring temporary ring artifact images Rt0, Rt1 or Rt1, Rt2 do not appear at the same position, but at a position rotated at an angle δ about the center of rotation Dz.

As illustrated in the method step C2 in FIG. 3, before the averaging the temporary ring artifact images Rt0, Rt1, Rt2 are rotated in such a way that the ring artifacts S0, S1, S2 come to lie at the same position within the image. The rotation is carried out in this case according to the following calculation rule and can, for example, be carried out on the basis of images illustrated in polar coordinates:

R k = l = 0 2 w l · Rot l · δ ( Rt l + k )
where l is an index traversing the values 0 to 2, w1 is a weighting factor dependent on the index l, δ is the angular spacing of ring artifacts between two neighboring tomograms, and Rot is a rotation operator that rotates the temporary ring artifact image Rt1+k by the angle 1*δ about the center of rotation, the following relationships holding:

l = 0 2 w l = 1
and δ=2*n*d/V, and d corresponding to a difference between the neighboring recording positions and V corresponding to feed of the recording system in the direction of the system axis. The factor 1/3, for example, can in each case be used as weighting factors w0, w1, w2.

Following the calculation of the ring artifact image I1, instead of the low-pass filtering T1, illustrated in FIG. 2, in the temporary ring artifact images Rt0, Rt1, Rt2, it is possible to perform an alternative low-pass filtering T2, indicated by dashes in FIG. 2, in the respective final ring artifact image R1. The alternative low-pass filtering T2 in the ring artifact image R1 is undertaken in accordance with the low-pass filtering T1 in the temporary ring artifact image Rt0 or Rt1 or Rt2 in the azimuth direction along a plurality of circular segments that respectively correspond to a part of a circle seated at the center of rotation Dz.

In the fourth method step D, a subtraction is carried out between the tomogram I1 and the ring artifact image R1 such that a result image E1 is produced in each case in which the ring artifact is removed.

Any of the aforementioned methods may be embodied in the form of a system or device, including, but not limited to, any of the structure for performing the methodology illustrated in the drawings.

Further, any of the aforementioned methods may be embodied in the form of a program. The program may be stored on a computer readable media and is adapted to perform any one of the aforementioned methods when run on a computer device (a device including a processor). Thus, the storage medium or computer readable medium, is adapted to store information and is adapted to interact with a data processing facility or computer device to perform the method of any of the above mentioned embodiments.

The storage medium may be a built-in medium installed inside a computer device main body or a removable medium arranged so that it can be separated from the computer device main body. Examples of the built-in medium include, but are not limited to, rewriteable non-volatile memories, such as ROMs and flash memories, and hard disks. Examples of the removable medium include, but are not limited to, optical storage media such as CD-ROMs and DVDs; magneto-optical storage media, such as MOs; magnetism storage media, such as floppy disks (trademark), cassette tapes, and removable hard disks; media with a built-in rewriteable non-volatile memory, such as memory cards; and media with a built-in ROM, such as ROM cassettes.

Example embodiments being thus described, it will be obvious that the same may be varied in many ways. Such variations are not to be regarded as a departure from the spirit and scope of the present invention, and all such modifications as would be obvious to one skilled in the art are intended to be included within the scope of the following claims.

Claims

1. A method for removing ring artifacts from tomograms produced with the aid of a computed tomography unit including at least one recording system rotating about a rotation axis, the method comprising:

reconstructing N tomograms Ik (k=1,..., N) at recording positions that are substantially equidistant in the direction of the rotation axis;
calculating a temporary ring artifact image Rtk for each tomogram Ik which has at least one ring artifact present in the respective tomogram Ik;
calculating a ring artifact image Rk for each tomogram Ik, the respective ring artifact image Rk being formed by averaging over at least a portion Tsub of the temporary ring artifact images Rtk, and the temporary ring artifact images Rtk being rotated before averaging in such a way that the ring artifacts present in the temporary artifact images Rtk substantially coincide and their position substantially corresponds to the ring artifacts present in the tomogram Ik; and
subtracting the respective ring artifact image Rk from the respective tomogram Ik, in each case one result image Ek being produced in which the ring artifact is removed.

2. The method as claimed in claim 1, wherein the calculation of a temporary ring artifact image Rtk comprises:

masking bone and air fractions in the respective tomogram Ik such that a masked image Nk is respectively produced for each tomogram Ik,
subjecting the respective masked image Nk to high-pass filtering in the radial direction relative to the center of rotation in the masked image Nk such that a high-pass-filtered image Hk is produced for each masked image Nk, and
artifact threshold value formation in the respective high-pass-filtered image Hk with a negative artifact threshold and a positive artifact threshold, so as to produce for each high-pass-filtered image Hk the temporary ring artifact image Rtk that has the ring artifacts present in the respective tomogram Ik.

3. The method as claimed in claim 2, wherein the masking of bone and air fractions comprises:

setting all the image values greater than an upper threshold SWO equal to SWO, and
setting all the image values smaller than a lower threshold SWU equal to SWU such that a masked image Nk is produced.

4. The method as claimed in claim 2, wherein the high-pass filtering comprises:

carrying out median filtering in radial directions, running through the center of rotation, in the masked image Nk such that a median-filtered image Mk is produced, and
subtracting the median-filtered image Mk from the tomogram Ik such that a high-pass-filtered image Hk is produced.

5. The method as claimed in claim 1, wherein the method includes, before the calculation of the ring artifact image Rk:

carrying out low-pass filtering in the respective temporary ring artifact image Rtk in the azimuth direction along at least one circular segment that corresponds to a part of a circle seated at the center of rotation and which has a circular segment length matched to the ring artifact.

6. The method as claimed in claim 1, wherein after the calculation of the ring artifact image Rk, the method includes:

carrying out low-pass filtering in the ring artifact image Rk in the azimuth direction along at least one circular segment that corresponds to a part of a circle seated at the center of rotation and which has a circular segment length matched to the ring artifact.

7. The method as claimed in claim 1, wherein the calculation of the ring artifact image Rk is performed using the following calculation rule: R k = ∑ l = - T + T ⁢ w l · Rot l · δ ⁡ ( Rt l + k ), where l is an index traversing the values −T and +T, wl is a weighting factor dependent on the index l, δ is the angular spacing of ring artifacts between two neighboring tomograms, and Rot is a rotation operator that rotates the temporary ring artifact image Rtl+k by the angle 1*δ about the center of rotation, the following relationships holding: T sub = 2 * T + l, ∑ l = - T T ⁢ w l = 1 and δ=2*n*d/V, and d corresponding to a difference between the neighboring recording positions and V corresponding to feed of the recording system in the direction of the system axis.

8. The method as claimed in claim 7, wherein the following holds for the weighting factors wl:wl=1/Tsub.

9. The method as claimed in claim 1, wherein at least some of the method steps are carried out in polar coordinates with an origin of coordinates seated at the center of rotation of the images.

10. A computed tomography unit that is designed for carrying out the method as claimed in claim 1.

11. The method as claimed in claim 3, wherein the high-pass filtering comprises:

carrying out median filtering in radial directions, running through the center of rotation, in the masked image Nk such that a median-filtered image Mk is produced, and
subtracting the median-filtered image Mk from the tomogram Ik such that a high-pass-filtered image Hk is produced.

12. The method as claimed in claim 2, wherein the method includes, before the calculation of the ring artifact image Rk:

carrying out low-pass filtering in the respective temporary ring artifact image Rtk in the azimuth direction along at least one circular segment that corresponds to a part of a circle seated at the center of rotation and which has a circular segment length matched to the ring artifact.

13. The method as claimed in claim 3, wherein the method includes, before the calculation of the ring artifact image Rk:

carrying out low-pass filtering in the respective temporary ring artifact image Rtk in the azimuth direction along at least one circular segment that corresponds to a part of a circle seated at the center of rotation and which has a circular segment length matched to the ring artifact.

14. The method as claimed in claim 11, wherein the method includes, before the calculation of the ring artifact image Rk:

carrying out low-pass filtering in the respective temporary ring artifact image Rtk in the azimuth direction along at least one circular segment that corresponds to a part of a circle seated at the center of rotation and which has a circular segment length matched to the ring artifact.

15. The method as claimed in claim 5, wherein after the calculation of the ring artifact image Rk, the method includes:

carrying out low-pass filtering in the ring artifact image Rk in the azimuth direction along at least one circular segment that corresponds to a part of a circle seated at the center of rotation and which has a circular segment length matched to the ring artifact.

16. The method as claimed in claim 14, wherein after the calculation of the ring artifact image Rk, the method includes:

carrying out low-pass filtering in the ring artifact image Rk in the azimuth direction along at least one circular segment that corresponds to a part of a circle seated at the center of rotation and which has a circular segment length matched to the ring artifact.

17. The method as claimed in claim 2, wherein the calculation of the ring artifact image Rk is performed using the following calculation rule: R k = ∑ l = - T + T ⁢ w l · Rot l · δ ⁡ ( Rt l + k ), where l is an index traversing the values −T and +T, wl is a weighting factor dependent on the index l, δ is the angular spacing of ring artifacts between two neighboring tomograms, and Rot is a rotation operator that rotates the temporary ring artifact image Rtl+k by the angle 1*δ about the center of rotation, the following relationships holding: T sub = 2 * T + l, ∑ l = - T T ⁢ w l = 1 and δ=2*n*d/V, and d corresponding to a difference between the neighboring recording positions and V corresponding to feed of the recording system in the direction of the system axis.

18. The method as claimed in claim 17, wherein the following holds for the weighting factors wl:wl=1/Tsub.

19. A device for removing ring artifacts from tomograms produced with the aid of a computed tomography unit including at least one recording system rotating about a rotation axis, the device comprising:

means for reconstructing N tomograms Ik (k=1,..., N) at recording positions that are substantially equidistant in the direction of the rotation axis;
means for calculating a temporary ring artifact image Rtk for each tomogram Ik which has at least one ring artifact present in the respective tomogram Ik;
means for calculating a ring artifact image Rk for each tomogram Ik, the respective ring artifact image Rk being formed by averaging over at least a portion Tsub of the temporary ring artifact images Rtk, and the temporary ring artifact images Rtk being rotated before averaging in such a way that the ring artifacts present in the temporary artifact images Rtk substantially coincide and their position substantially corresponds to the ring artifacts present in the tomogram Ik; and
means for subtracting the respective ring artifact image Rk from the respective tomogram Ik, in each case one result image Ek being produced in which the ring artifact is removed.

20. A computer readable medium including program segments for, when executed on a computer, causing the computer to implement the method of claim 1.

Referenced Cited
U.S. Patent Documents
5127037 June 30, 1992 Bynum
5644610 July 1, 1997 Crawford et al.
5867554 February 2, 1999 Hupke
6031374 February 29, 2000 Epstein et al.
6047039 April 4, 2000 Flohr
6448559 September 10, 2002 Saoudi et al.
7053376 May 30, 2006 Amemiya et al.
Foreign Patent Documents
198 35 451 March 1999 DE
0 819 406 January 1998 EP
Patent History
Patent number: 7391927
Type: Grant
Filed: Sep 14, 2005
Date of Patent: Jun 24, 2008
Patent Publication Number: 20060056579
Assignee: Siemens Aktiengesellschaft (Munich)
Inventor: Karl Stierstorfer (Erlangen)
Primary Examiner: Yosef Kassa
Attorney: Harness, Dickey & Pierce, P.L.C.
Application Number: 11/225,035